1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
|
#! /usr/bin/env python
# Matthew Wakefield Feb 2004
from __future__ import with_statement
import unittest
import os
import warnings
from cogent import LoadSeqs, LoadTree
import cogent.evolve.parameter_controller, cogent.evolve.substitution_model
from cogent.maths import optimisers
__author__ = "Peter Maxwell"
__copyright__ = "Copyright 2007-2012, The Cogent Project"
__credits__ = ["Peter Maxwell", "Gavin Huttley", "Matthew Wakefield"]
__license__ = "GPL"
__version__ = "1.5.3"
__maintainer__ = "Gavin Huttley"
__email__ = "gavin.huttley@anu.edu.au"
__status__ = "Production"
base_path = os.getcwd()
data_path = os.path.join(base_path, 'data')
good_rule_sets = [
[
{'par_name' : 'length','is_independent':True},
],
[
{'par_name' : 'length','is_independent':True},
],
[
{'par_name' : 'length','is_clade' :True, 'is_independent':True, 'edges' : ['a','b']},
],
[
{'par_name' : 'length','is_independent':True, 'edges' : ['a','c','e']},
],
[
{'par_name' : 'length','is_independent':True, 'edge' : 'a'},
],
]
bad_rule_sets = [
[
{'par_name' : 'length','is_clade' :True, 'edges' : ['b','f'],},
],
]
class test_parameter_controller(unittest.TestCase):
"""Tesing Parameter Controller"""
def setUp(self):
#length all edges 1 except c=2. b&d transitions all other transverions
self.al = LoadSeqs(
data={'a':'tata', 'b':'tgtc', 'c':'gcga', 'd':'gaac', 'e':'gagc',})
self.tree = LoadTree(treestring='((a,b),(c,d),e);')
self.model = cogent.evolve.substitution_model.Nucleotide(
do_scaling=True, equal_motif_probs=True, model_gaps=True)
def test_scoped_local(self):
model = cogent.evolve.substitution_model.Nucleotide(
do_scaling=True, equal_motif_probs=True, model_gaps=True,
predicates = {'kappa':'transition'})
lf = model.makeLikelihoodFunction(self.tree)
lf.setConstantLengths()
lf.setAlignment(self.al)
null = lf.getNumFreeParams()
lf.setParamRule(par_name='kappa',
is_independent=True,
edges=['b','d'])
self.assertEqual(null+2, lf.getNumFreeParams())
def test_setMotifProbs(self):
"""Mprobs supplied to the parameter controller"""
model = cogent.evolve.substitution_model.Nucleotide(
model_gaps=True, motif_probs=None)
lf = model.makeLikelihoodFunction(self.tree,
motif_probs_from_align=False)
mprobs = {'A':0.1,'C':0.2,'G':0.2,'T':0.5,'-':0.0}
lf.setMotifProbs(mprobs)
self.assertEqual(lf.getMotifProbs(), mprobs)
lf.setMotifProbsFromData(self.al[:1], is_constant=True)
self.assertEqual(lf.getMotifProbs()['G'], 0.6)
lf.setMotifProbsFromData(self.al[:1], pseudocount=1)
self.assertNotEqual(lf.getMotifProbs()['G'], 0.6)
# test with consideration of ambiguous states
al = LoadSeqs(data = {'seq1': 'ACGTAAGNA', 'seq2': 'ACGTANGTC',
'seq3': 'ACGTACGTG'})
lf.setMotifProbsFromData(al, include_ambiguity=True, is_constant=True)
motif_probs = dict(lf.getMotifProbs())
correct_probs = {'A': 8.5/27, 'C': 5.5/27, '-': 0.0, 'T': 5.5/27,
'G': 7.5/27}
self.assertEqual(motif_probs, correct_probs)
self.assertEqual(sum(motif_probs.values()), 1.0)
def test_setMultiLocus(self):
"""2 loci each with own mprobs"""
model = cogent.evolve.substitution_model.Nucleotide(motif_probs=None)
lf = model.makeLikelihoodFunction(self.tree,
motif_probs_from_align=False, loci=["a", "b"])
mprobs_a = dict(A=.2, T=.2, C=.3, G=.3)
mprobs_b = dict(A=.1, T=.2, C=.3, G=.4)
for is_constant in [False, True]:
lf.setMotifProbs(mprobs_a, is_constant=is_constant)
s = str(lf)
lf.setMotifProbs(mprobs_b, locus="b")
self.assertEqual(lf.getMotifProbs(locus="a"), mprobs_a)
self.assertEqual(lf.getMotifProbs(locus="b"), mprobs_b)
s = str(lf)
#lf.setParamRule('mprobs', is_independent=False)
def test_setParamRules(self):
lf = self.model.makeLikelihoodFunction(self.tree)
def do_rules(rule_set):
for rule in rule_set:
lf.setParamRule(**rule)
for rule_set in good_rule_sets:
lf.setDefaultParamRules()
do_rules(rule_set)
for rule_set in bad_rule_sets:
lf.setDefaultParamRules()
self.assertRaises((KeyError, TypeError,
AssertionError, ValueError), do_rules, rule_set)
def test_setLocalClock(self):
pass
def test_setConstantLengths(self):
t = LoadTree(treestring='((a:1,b:2):3,(c:4,d:5):6,e:7);')
lf = self.model.makeLikelihoodFunction(t)#self.tree)
lf.setParamRule('length', is_constant=True)
# lf.setConstantLengths(t)
lf.setAlignment(self.al)
self.assertEqual(lf.getParamValue('length', 'b'), 2)
self.assertEqual(lf.getParamValue('length', 'd'), 5)
def test_pairwise_clock(self):
al = LoadSeqs(data={'a':'agct','b':'ggct'})
tree = LoadTree(treestring='(a,b);')
model = cogent.evolve.substitution_model.Dinucleotide(
do_scaling=True, equal_motif_probs=True, model_gaps=True,
mprob_model='tuple')
lf = model.makeLikelihoodFunction(tree)
lf.setLocalClock('a','b')
lf.setAlignment(al)
lf.optimise(local=True)
rd = lf.getParamValueDict(['edge'], params=['length'])
self.assertAlmostEquals(lf.getLogLikelihood(),-10.1774488956)
self.assertEqual(rd['length']['a'],rd['length']['b'])
def test_local_clock(self):
lf = self.model.makeLikelihoodFunction(self.tree)
lf.setLocalClock('c','d')
lf.setAlignment(self.al)
lf.optimise(local=True,
tolerance=1e-8, max_restarts=2)
rd = lf.getParamValueDict(['edge'], params=['length'])
self.assertAlmostEquals(lf.getLogLikelihood(),-27.84254174)
self.assertEqual(rd['length']['c'],rd['length']['d'])
self.assertNotEqual(rd['length']['a'],rd['length']['e'])
def test_complex_parameter_rules(self):
# This test has many local minima and so does not cope
# with changes to optimiser details.
model = cogent.evolve.substitution_model.Nucleotide(
do_scaling=True, equal_motif_probs=True, model_gaps=True,
predicates = {'kappa':'transition'})
lf = model.makeLikelihoodFunction(self.tree)
lf.setParamRule(par_name='kappa',
is_independent=True)
lf.setParamRule(par_name='kappa',
is_independent=False,
edges=['b','d'])
lf.setConstantLengths(LoadTree(
treestring='((a:1,b:1):1,(c:2,d:1):1,e:1);'))
#print self.pc
lf.setAlignment(self.al)
lf.optimise(local=True)
rd = lf.getParamValueDict(['edge'], params=['kappa'])
self.assertAlmostEquals(lf.getLogLikelihood(),-27.3252, 3)
self.assertEqual(rd['kappa']['b'],rd['kappa']['d'])
self.assertNotEqual(rd['kappa']['a'],rd['kappa']['b'])
def test_bounds(self):
"""Test setting upper and lower bounds for parameters"""
lf = self.model.makeLikelihoodFunction(self.tree)
lf.setParamRule('length', value=3, lower=0, upper=5)
# Out of bounds value should warn and keep bounded
with warnings.catch_warnings(record=True) as w:
lf.setParamRule('length', lower=0, upper=2)
self.assertTrue(len(w), 'No warning issued')
self.assertEqual(lf.getParamValue('length', edge='a'), 2)
# upper < lower bounds should fail
self.assertRaises(ValueError, lf.setParamRule,
'length', lower=2, upper=0)
if __name__ == '__main__':
unittest.main()
|
